agentcache

Make Your AI Agents Lightning Fast

The caching layer that makes Puppeteer & Playwright agents 25-80% faster and 50% cheaper. One line of code. Zero configuration.

500+ agents, 10M+ network requests/s, and 10K+ cache hits/s in prod

Cost calculator

1,000
1K runs (~4GB)
≤$2.00
1M requests, ~2K RPM (~7GB)
≤$5.00
1M queries, ~2K QPS (~338TB)
100K agents
per agent
Estimated cost
$9.69/month
(min $64/month for launch plan)
Workload768 dimensions, 1M docs, ~3GB
p508ms
343ms
p9018ms
444ms
p9935ms
554ms
Warm namespace
Cold namespace
Approach (32 QPS with topk=10)

One-line integration

Automatically cache all network requests in your Playwright/Puppeteer agents

import { attachAgentCache } from 'agent-cache'

const context = await browser.newContext()
attachAgentCache(context) // ✨ That's it!

// All network requests are now cached automatically
await page.goto('https://example.com') // Cached on repeat visits
await page.click('button') // API calls cached across runs
25-80%
Faster Performance

Eliminate redundant network requests across agent runs

50%
Cost Reduction

Cut bandwidth, proxy and egress costs

1-Line
SDK Integration

Simple attachAgentCache() call

Ready to accelerate your agents?

Join 500+ developers using AgentCache to build faster, more cost-effective AI agents.

14-day free trial • No setup fees • Cancel anytime